The Red Queen Effect

The Red Queen Effect describes competitive environments where participants must continuously invest and innovate just to maintain their current position — where standing still is equivalent to falling behind. The name comes from Lewis Carroll's Through the Looking-Glass, in which the Red Queen tells Alice: "Now, here, you see, it takes all the running you can do, to keep in the same place."

Evolutionary Origins

The concept was formalized by evolutionary biologist Leigh Van Valen in 1973 as the Red Queen Hypothesis. Van Valen observed that the probability of extinction for a species appeared roughly constant over time — meaning that species never truly "get ahead" in the evolutionary arms race. Predators evolve faster jaws; prey evolve thicker shells. Parasites evolve to exploit hosts; hosts evolve resistance. The fitness landscape is always shifting because every organism's competitors are also evolving. Absolute fitness can increase, but relative fitness — the only kind that matters for survival — remains a treadmill.

This insight extends beyond predator-prey dynamics. Sexual selection itself may be a Red Queen phenomenon: organisms that reproduce sexually pay an enormous cost (only half their genes pass to each offspring) but gain the ability to shuffle genetic combinations rapidly enough to stay ahead of co-evolving parasites. The Red Queen explains why sex exists despite its apparent inefficiency.

The Red Queen in Competitive Markets

In economics, the Red Queen Effect manifests wherever competitive position depends on relative advantage rather than absolute capability. The dynamic is especially brutal in technology markets, where innovation by one firm endogenously depreciates the value of rivals' existing investments.

Consider the foundation model race. When OpenAI releases a more capable model, it doesn't just improve OpenAI's position — it actively degrades the competitive value of every other lab's current model. Anthropic, Google DeepMind, and Meta must respond not because their models got worse in absolute terms, but because the market's expectations shifted. A model that was state-of-the-art six months ago is now merely adequate. This is why AI labs are spending billions per training run — they're not building durable assets so much as buying temporary position in a race where the finish line moves forward every quarter.

Zhang and Zhang's 2026 paper The Economics of Digital Intelligence Capital formalizes this as endogenous depreciation: the value of a foundation model depreciates not through wear or obsolescence in the traditional sense, but because competitors' innovations redefine the capability frontier. Unlike physical capital, which depreciates on a predictable schedule, AI capital depreciates on a schedule set by your competitors' R&D velocity. This makes the economics of AI uniquely punishing — and uniquely demanding of continuous investment.

Red Queen Dynamics in the Agentic Economy

The Red Queen operates at every layer of the agentic economy. At the silicon layer, NVIDIA must ship new GPU architectures on an annual cadence — its $26 billion commitment to training its own foundation models is itself a Red Queen move, ensuring it doesn't cede the model layer to customers who might eventually design around its hardware. At the platform layer, cloud providers continuously expand inference capabilities and cut prices because any pause lets competitors capture developers. At the application layer, AI-native products must continuously integrate new model capabilities or watch users migrate to alternatives that do.

The Red Queen also explains why the AI industry's massive capital expenditure doesn't lead to the overcapacity that skeptics predict. In a Red Queen market, today's infrastructure isn't competing against nothing — it's competing against the infrastructure that will exist in twelve months. Companies invest not because current demand justifies it, but because falling behind in capability or scale creates a competitive deficit that compounds. The Jevons Paradox amplifies this further: each round of Red Queen investment produces efficiency gains that expand total demand, financing the next round of investment.

The deepest Red Queen dynamic may be in talent. As AI tools make individual engineers more productive, the bar for what constitutes competitive output rises for everyone. Engineers who adopt AI coding tools don't get to relax — they're expected to produce proportionally more, because their competitors are doing the same. Productivity gains don't create leisure; they reset the baseline. The Red Queen runs through organizations, not just between them.